CyberAgentAILab/VLG-Loc-project-page

Vision-Language Global Localization (VLG-Loc) is a global localization method that uses camera images and a human-readable labeled footprint map containing only names and areas of distinctive visual landmarks